Is Micron Stock a Buy as Revenue Continues to Surge?

Micron reported fiscal Q3 revenue of $41.5B (from $9.3B), beating the $35.8B consensus, with gross margins surging to 84.6% (from 37.7% a year ago) and adjusted EPS of $24.67 vs $20.78 expected. Guidance calls for fiscal Q4 revenue of ~$50B and gross margins ~86%, with adjusted EPS of ~$30.73 at the midpoint, supported by AI-driven DRAM/NAND tight supply and capex raised to $27B. Long-term customer agreements (~40% of revenue) and HBM supply booked through 2028 reinforce earnings visibility despite cyclical risk.

Analysis

MU is the clearest direct beneficiary of AI infrastructure intensity, but the bigger market mechanism is not just a better quarter — it is a pricing regime shift. When a component becomes booked out multiple years and customers sign more take-or-pay style agreements, the equity stops trading like a pure cyclical and starts trading like a constrained-capacity utility with cyclical upside. That supports multiple expansion near term, but also raises the odds that investors are underestimating how much of the current economics will be competed away once new wafer capacity and HBM lines come online.

The second-order winner set is the AI stack itself: NVDA and hyperscaler capex budgets still benefit from AI demand, but tighter memory availability can cap system shipments and keep GPU BOM costs elevated, which may compress gross margins at the margin for integrators and cloud buyers. The losers are downstream server OEMs, OEMs with weaker supply agreements, and any AI-exposed hardware name that cannot pass through memory inflation. Over 6-18 months, MU’s own capex response is the key risk: the market is effectively pricing scarcity, but new greenfield spend is how scarcity ends.

The near-term catalyst is confirmation that forward bookings and pricing power persist into the next two quarters; the reversal trigger is any hint of contract normalization, lead-time shortening, or a capex guide that implies supply catches up faster than expected. Contrarian view: the stock may not be as cheap as the headline P/E suggests if earnings are near peak and the market is capitalizing unsustainably high margins. Consensus is likely right on the next 1-2 quarters, but may be overconfident on 2027 earnings durability.

MU is the best outright long on a 1-3 month horizon, but we would pair it against a basket of AI hardware beneficiaries that are more exposed to input-cost inflation. The cleaner relative trade is long MU / short a semiconductor ETF or weaker memory-dependent hardware names; if MU loses >15% from current levels after a guide miss or memory prices stop rising, the thesis should be revisited.

Key Decisions for Investors

• Long MU for 1-3 months on post-earnings strength; use any 5-8% pullback to add. Thesis: book-out visibility and pricing power can drive another 10-20% rerating if next guide holds.
• Pair trade: long MU / short SOXX or a semis basket for 1-2 quarters. This isolates memory scarcity versus broader AI multiple risk; invalidate if MU lead times compress or gross margin guidance rolls over.
• Buy MU call spreads 3-6 months out rather than stock if chasing momentum. Risk/reward works only if the market continues to pay for scarcity; max loss should be capped if capex overshoots and the cycle normalizes.
• Watch NVDA as a second-order beneficiary but not the primary expression. If memory inflation begins to constrain GPU shipments, take profits on NVDA strength and rotate toward MU.
• Falsifier alert: if MU guides for materially lower gross margin or capex-driven supply growth starts to outpace demand commentary in the next 1-2 quarters, reduce exposure aggressively.